It only took 24 hours of complaints to Openai to start bringing back his old model

Openai unveiled its latest generator of generative AI, GPT-5 on Thursday. CEO Sam Altman says that Chatgpt is now like having a “superpower” and the equivalent of a “legitimate expert at the doctorate level in all, in any field you need, on demand, which can help you with your goals”. But after a day of play with it, many people are disappointed. Not only because GPT-5 still escapes basic questions, but because it seems to break a lot of workflows, according to complaints published on Reddit.

How many people do they hate what happened with GPT-5? Altman now says that they bring back the last model for paid users. “We will let users choose to continue using 4O. We will look at the use by thinking of the duration to offer inherited models, ”tweeted Altman.

Altman also wrote that the company would double the GPT-5 rate limits for Chatgpt Plus users and blamed the fact that the new model seemed “more stupid” on the rupture of the Autoswitcher. The CEO also said that they were going to change the user interface to facilitate switching between different models. “Everyone’s deployment takes a little more time. It is a massive change on a large scale. For example, our API traffic has dubbed over the past 24 hours …”, wrote Altman.

Chatgpt users are seriously upset and it is not at all clear, however, if Altman’s promises will compensate for this. Because it’s not just 4o for which people claim.

“I woke up this morning to see that Optai deleted 8 models overnight. No warning. No choice. No “inherited option”. They just deleted them, ”complained an R / Chatgpt user. “4o? Gone. O3? Gone. O3-Pro? Gone. 4.5? Gone. Anything that made chatgpt really useful for my workflow – deleted. “

The user wrote that 4o was not only a tool for them: “It helped me to go through anxiety, depression and some of the darkest periods of my life. He had this warmth and this understanding that felt … human. “

Another user on R / Chatgpt complained that he had the impression that he was now using a free version with GPT-5 despite being a paid subscriber: “I am so disappointed, just like the millions of people here. A company that manages the largest AI model cannot understand what its users want. The biggest pen (sic) shit in industry. ”

Another Reddit user explained again why they used different models and how to deactivate them was devastating, explaining that they had now canceled their paid subscription after two years:

What type of company removes a workflow from 8 models Over the next day, without prior warning to their paid users?

I don’t think I have to talk to myself when I say that each model was useful for a specific use case (all the logic behind several models with variable capabilities). Essentially distribute your workflow in several agents with specific tasks.

Personally, 4o was used for creativity and emerging ideas, O3 was used for pure logic, O3-Pro for deep research, 4.5 for writing, etc. I’m sure many of you have experienced the same type of thing.

The user pursued by thinking that there was a harmful goal behind the switch, floating that he was part of a conspiracy theory to delete creativity: “Openai drafs users to believe that this removal engine is the “most intelligent model on earth”, simultaneously removing the models that showed real emergence and creativity. The user even used the term “social control”, strongly leaning on the idea that the dark forces were preparing for “societal collapse”.

Other commentators on boards outside of R / Chatgpt have seen it less as a sign of collapse or control of society. They simply supposed that the movements have proven that the emperor did not wear clothes. An R / Technology user wrote: “The Chatgpt bubble jumped today with the gravity of these SAM lies. He has lost all confidence in the future.”

Even if you ignore workflow problems (and you really shouldn’t), the GPT-5 is still far from perfect. People spent the day on social media platforms like Bluesky producing the most stupid examples of Chatppt that was lease.

Altman recognized on X that its deployment did not go well. “We will continue to work so that stable things and will continue to listen to the comments,” tweeted the CEO of Openai. “As we mentioned, we expected a little bump by unrolling so much at the same time. But it was a little more joyful than we were hoping for! ”

It could be difficult for Chatgpt to recover, especially since so many people on Reddit claim that they cancel their subscriptions. And Openai has many competitors like Claude d’Anthropic, Grok de Xai and Google’s Gemini. But we should find out early enough if the 4O radiation is enough for most chatgpt users. Otherwise, they probably jump from the ship.